//context
I joined Mondly after Pearson acquired it to help redesign the iOS and Android app for 1M+ monthly learners. I inherited an initial concept and, after a two-month handover, spent nearly a year as the only designer alongside around 40 engineers.
I owned the flows, interaction details, research, and design QA through implementation. As the work grew, I hired four designers and established the design-system foundation and research practice. The redesign improved lesson completion and conversion; the figures below describe the team's results.
Monthly active users
1M+
Lesson completion rate
+18%
Conversion rate
+32%
Design team scaled
1→5
//redesigning the core journey
The app needed more than an updated appearance. Onboarding asked people to work through unnecessary steps before they understood the experience, while features such as Daily Lessons and Chatbot were easy to miss inside the map.
I took the existing redesign direction through the flows and details needed for implementation. I simplified onboarding, brought learning activities into the Explore page, and made the next lesson a clear action on the map. The aim was to help people get started and know where to go next, while keeping familiar functionality intact.
With a legacy codebase and one designer supporting around 40 engineers, I had to prioritise. I focused on those core journeys and worked through platform differences and edge cases with engineering as we built.
//improving learning after launch
Once the redesign shipped, my focus shifted to what happened inside lessons: helping learners practise, understand feedback, and revisit vocabulary. We used A/B tests and iterative design to improve lesson and category completion.
We adapted more than 25 quiz types for small screens, low-end devices, and right-to-left scripts such as Arabic. Motion, haptics, and audio helped learners follow instructions and recognise feedback.
Category completion rate
+100%
Quiz types modernized
25+
Flashcard feature
Spaced-repetition flashcards help learners revisit words from their lessons.
listen & repeat exercise
Learners hear a phrase and practise saying it before being tested. This addressed feedback that lessons jumped straight to assessment.
course map
An interactive course map shows learners where they are and which lesson comes next.
RTL support
Layout guidelines for right-to-left scripts, including Arabic, within the multilingual app.
Lesson review
Layered Lottie animations create a playful review moment with the globe mascot.
AI conversation
Concept: mission-based conversation practice. Learners reply to AI prompts, get suggested phrases when stuck, and pick formal or informal tone.
//Vocabulary Review
Learners wanted a way to revisit words outside a lesson. Our first release let them save a lesson's words into one folder.
We then added saving within lessons and explored custom collections, checking whether people found the feature and returned to use it.
MVP
The first release let learners review a lesson’s words and save them in one folder.
In-lesson save to collection feature
Learners can save a word as they encounter it, without leaving the lesson.
custom collections
Concept: custom collections that turn a flat saved list into named, personal study sets.
//research
I introduced research alongside the redesign: a survey sent to 20,000 new and churned learners, interviews, usability tests, and live product feedback. Complaints about being tested before practising, for example, informed the listen-and-repeat exercise.
//design system
I built the design-system foundation while the redesign was in development. Shared semantic tokens now support consistent styling across the app. A full reusable component system is still unfinished.
//Interactions & motion design
I designed transitions, haptics, and audio cues across onboarding, lessons, and review. The globe mascot and lesson-end animations gave learners a clear moment of completion.
//challenges of tech debt
The legacy codebase made changes slow and left gaps between iOS and Android. I prioritised the most disruptive usability problems, kept familiar navigation patterns, and worked with engineers on what each platform could support.
